Data Amount Heterogeneity Improves Performance in Federated Learning for Neuroimaging

Paul Kurian; Tamoghna Chattopadhyay; Abhijith Shaji; Greg Ver Steeg; Paul M. Thompson; Jose Luis Ambite

cs.LGOct 28, 2026

Data Amount Heterogeneity Improves Performance in Federated Learning for Neuroimaging

Authors: Paul Kurian, Tamoghna Chattopadhyay, Abhijith Shaji, Greg Ver Steeg, Paul M. Thompson, Jose Luis Ambite

Organizations: Information Sciences Institute, University of Southern California, Marina del Rey, CA, USA · Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Marina del Rey, CA, USA · University of California, Riverside, CA, USA

Abstract

Federated Learning (FL) allows for secure analysis of distributed data without data sharing. We and others have successfully applied FL to neuroimaging tasks, such as Alzheimer's disease classification from structural magnetic resonance imaging (MRI). FL trains a deep learning model locally at each site, encrypts the local model parameters and sends them to a federation controller that aggregates the local models under homomorphic encryption, the resulting community model is sent back to the sites for further training, and the process repeats for a number of rounds. FL often achieves learning performance (e.g., accuracy) close to centralized training, but no data leaves the sites and model parameters are protected through encryption and other measures. In this paper, we show that federations composed of sites with heterogeneous amounts of data (1 large site plus small sites) perform better than homogeneous (same data per site) federations, particularly in low data regimes, while reaching performance close to centralized training. Having a big site often results in faster convergence and higher final accuracy. Heterogeneous federations better exploit sites with small amounts of data, a common situation in large consortia such as ENIGMA.

Figures & tables

Fig. 1 : 3D Convolutional Neural Network for binary classification of AD status from T1 MRI.

Fig. 2 : Centralized performance (accuracy) of AD classification for different amounts of data (average over 5 runs, shaded band of 1 standard deviation).

Panel of Figure 3 — Fig. 3 : Federated Learning performance (AUC) of AD classification for different amounts of data for centralized, homogeneous, and heterogeneous environments (average over 10 runs, shaded band of 1 standard deviation).

Panel of Figure 4 — Fig. 4 : Federated Learning performance (Accuracy) of AD classification for different amounts of data for centralized, homogeneous and heterogeneous environments (average over 10 runs, shaded band of 1 standard deviation).

Panel of Figure 5 — Fig. 5 : Federated Learning performance, Accuracy and AUC (ROC), of AD classification for different amounts of data for centralized, homogeneous and heterogeneous environments. Average test accuracy over 5 runs using early stopping, shaded band of 1 standard deviation.

Federated Learning

Federated learning trains shared models across data holders without requiring them to pool their raw data.

See all 961 papers

Evaluating Federated Learning approaches for mammography under breast density heterogeneity
May 9, 2026Gonzalo Iñaki Quintana, Franco Martin Di Maria, Laurence VancambergFederated Learning Aggregation Medical Image Analysis
GE HealthCare, 283 Rue de la Mini`ere, 78530 Buc, France · ENS Paris-Saclay, 4 Av. des Sciences, 91190 Gif-sur-Yvette, France · IMT Atlantique, 655 Av. du Technopˆole, 29280 Plouzan´e, France +1
Centralized vs Decentralized Federated Learning: A trade-off performance analysis
May 15, 2026Chaimaa Medjadji, Guilain Leduc, Sylvain Kubler +1Decentralized Learning Federated Learning
University of Luxembourg
A generalizable structural brain MRI foundation model via dual-priority federated pretraining
Sep 23, 2026Zhen Yu, Yang Liu, Jianfeng Feng +2Medical Imaging Foundation Models Federated Learning Aggregation
National Institute of Health Data Science, Peking University, Beijing, 100871, China. · Institute of Medical Technology, Peking University Health Science Center, Beijing, 100871, China. · State Key Laboratory of General Artificial Intelligence, Peking University, Beijing, 100871, China. +2

May 9, 2026J/K move · Enter open · S save

Evaluating Federated Learning approaches for mammography under breast density heterogeneity

Gonzalo Iñaki Quintana, Franco Martin Di Maria, Laurence Vancamberg

GE HealthCare, 283 Rue de la Mini`ere, 78530 Buc, France · ENS Paris-Saclay, 4 Av. des Sciences, 91190 Gif-sur-Yvette, France · IMT Atlantique, 655 Av. du Technopˆole, 29280 Plouzan´e, France+1

Federated Learning Aggregation Medical Image Analysis Non-IID Federated Learning Medical Image Classification Mammography Image Classification Federated Learning

Open paper AskPDF arXiv

Breast density is a key factor that influences mammography interpretation and is a major source of heterogeneity in multicenter datasets. Such heterogeneity poses challenges for collaborative machine learning across institutions, particularly in Federated Learning. This study aims to evaluate the impact of breast density-induced heterogeneity on FL for mammography image classification and to assess the robustness of common FL algorithms in realistic clinical settings. We conducted experiments under two scenarios: (1) a strongly heterogeneous setting where each participating site contributed exclusively low- or high-density cases, based on the BI-RADS density score, and (2) a population-based setting simulating breast density distributions in White and Asian populations. For the strongly heterogeneous setting, we evaluated two configurations: one with 2 clients, where the cases were grouped as BI-RADS A-B and C-D, and one with 4 clients, where each site contained cases of a single BI-RADS density. We compared three FL methods (FedAvg, FedProx, SCAFFOLD) against centralized training, local-only training, and naive aggregation approaches, including ensembling and weight averaging. Across both scenarios, FL achieved performance comparable to centralized training, while local models and naive aggregation approaches underperformed in the presence of strong heterogeneity. Notably, FedAvg achieved accuracy on par with or exceeding centralized training, demonstrating resilience to breast density-induced data imbalance without requiring specialized heterogeneity mitigation algorithms. These findings show that FL can address breast density-related heterogeneity, supporting its feasibility for real-world mammography workflows. The demonstrated robustness of FedAvg underscores the potential for broad clinical deployment of FL, enabling collaborative model development while maintaining data privacy.

Abstract

Federated Learning

Explore similar work

Evaluating Federated Learning approaches for mammography under breast density heterogeneity

Centralized vs Decentralized Federated Learning: A trade-off performance analysis

A generalizable structural brain MRI foundation model via dual-priority federated pretraining